arXiv subjects

Wenjie Wu

Publications and source records attributed to Wenjie Wu.

36 records · Page 2Linked to original sources

arXiv2025

GeoX: Geometric Problem Solving Through Unified Formalized Vision-Language Pre-training

Despite their proficiency in general tasks, Multi-modal Large Language Models (MLLMs) struggle with automatic Geometry Problem Solving (GPS), which demands understanding diagrams, interpreting symbols, and performing complex reasoning. This limitation arises from their pre-training on natural images and texts, along with the lack of automated verification in the problem-solving process. Besides, current geometric specialists are limited by their task-specific designs, making them less effective for broader geometric problems. To this end, we present GeoX, a multi-modal large model focusing on geometric understanding and reasoning tasks. Given the significant differences between geometric diagram-symbol and natural image-text, we introduce unimodal pre-training to develop a diagram encoder and symbol decoder, enhancing the understanding of geometric images and corpora. Furthermore, we introduce geometry-language alignment, an effective pre-training paradigm that bridges the modality gap between unimodal geometric experts. We propose a Generator-And-Sampler Transformer (GS-Former) to generate discriminative queries and eliminate uninformative representations from unevenly distributed geometric signals. Finally, GeoX benefits from visual instruction tuning, empowering it to take geometric images and questions as input and generate verifiable solutions. Experiments show that GeoX outperforms both generalists and geometric specialists on publicly recognized benchmarks, such as GeoQA, UniGeo, Geometry3K, and PGPS9k.

cs.CV↗

arXiv2024

The evolution of the mass--radius relation of expanding very young star clusters

The initial mass--radius relation of embedded star clusters is an essentially important boundary condition for understanding the evolution of embedded clusters in which stars form to their release into the galactic field via an open star cluster phase. The initial mass--radius relation of embedded clusters deduced by Marks \& Kroupa is significantly different from the relation suggested by Pfalzner et al. Here, direct N-body simulations are used to model the early expansion of embedded clusters after the expulsion of their residual gas. The observational data of radii of up to a few~Myr old clusters collated from various sources are found to be well fit by the N-body models, implying these observed very young clusters to most likely be in an expanding state. We show that the mass-radius relation of Pfalzner et al. reflects the expansion of embedded clusters following the initial mass-radius relation of Marks \& Kroupa. We also suggest that even the embedded clusters in ATLASGAL clumps with HII regions are probably already in expansion. All here collected clusters from different observations show a mass-size relation with a similar slope, which may indicate that all clusters were born with a profile resembling that of the Plummer phase-space distribution function.

astro-ph.GA↗

arXiv2024

DocGenome: An Open Large-scale Scientific Document Benchmark for Training and Testing Multi-modal Large Language Models

Scientific documents record research findings and valuable human knowledge, comprising a vast corpus of high-quality data. Leveraging multi-modality data extracted from these documents and assessing large models' abilities to handle scientific document-oriented tasks is therefore meaningful. Despite promising advancements, large models still perform poorly on multi-page scientific document extraction and understanding tasks, and their capacity to process within-document data formats such as charts and equations remains under-explored. To address these issues, we present DocGenome, a structured document benchmark constructed by annotating 500K scientific documents from 153 disciplines in the arXiv open-access community, using our custom auto-labeling pipeline. DocGenome features four key characteristics: 1) Completeness: It is the first dataset to structure data from all modalities including 13 layout attributes along with their LaTeX source codes. 2) Logicality: It provides 6 logical relationships between different entities within each scientific document. 3) Diversity: It covers various document-oriented tasks, including document classification, visual grounding, document layout detection, document transformation, open-ended single-page QA and multi-page QA. 4) Correctness: It undergoes rigorous quality control checks conducted by a specialized team. We conduct extensive experiments to demonstrate the advantages of DocGenome and objectively evaluate the performance of large models on our benchmark.

cs.CV↗

arXiv2024

MathLearner: A Large Language Model Agent Framework for Learning to Solve Mathematical Problems

With the development of artificial intelligence (AI), large language models (LLM) are widely used in many fields. However, the reasoning ability of LLM is still very limited when it comes to mathematical reasoning. Mathematics plays an important role in all aspects of human society and is a technical guarantee in the fields of healthcare, transport and aerospace, for this reason, the development of AI big language models in the field of mathematics has great potential significance. To improve the mathematical reasoning ability of large language models, we proposed an agent framework for learning to solve mathematical problems based on inductive reasoning. By emulating the human learning process of generalization of learned information and effective application of previous knowledge in new reasoning tasks, this framework has great performance in the mathematical reasoning process. It improves global accuracy over the baseline method (chain-of-thought) by 20.96% and solves 17.54% of the mathematical problems that the baseline cannot solve. Benefiting from the efficient RETRIEVAL method, our model improves the ability of large language models to efficiently use external knowledge, i.e., the mathematical computation of the model can be based on written procedures. In education, our model can be used as a personalised learning aid, thus reducing the inequality of educational resources.

cs.CL↗

arXiv2024

Open star clusters and their asymmetrical tidal tails

Stars that evaporate from their star cluster by the energy equipartition process end up either in a leading or a trailing tidal tail. In Newtonian gravitation and for open star clusters in the Solar vicinity, the tidal threshold, or prah, for escape is symmetrical, such that the leading and trailing tails are equally populated. The data by six independent teams that applied the convergent point method to map out the tidal tails of four open clusters (the Hyades, the Praesepe, Coma Berenices and COIN-Gaia13) using Gaia DR2 and DR3 are here applied to test for the expected symmetry. All tidal tails contain more stars in the leading tail. The combined confidence amounts to an 8 sigma falsification of the prah symmetry. The same test using Milgromian dynamics leads to consistency with the data. More effort needs to be exerted on this matter, but the data indicate with high confidence that the tidal prah of an open star cluster is asymmetrical with the corresponding confidence that Newtonian gravitation is falsified. Open star clusters depopulate more rapidly in Milgromian than in Newtonian dynamics and the COIN-Gaia13 cluster is here found to be nearly completely dissolved. In view of these results, the wide-binary star test and the Keplerian Galactic rotation curve finding are briefly discussed.

astro-ph.GA↗

arXiv2024

Mass segregation and velocity dispersion as evidence for a dark star cluster

A dark star cluster (DSC) is a system in which the cluster potential is dominated by stellar remnants, such as black holes and neutron stars having larger masses than the long-lived low-mass stars. Due to mass segregation, these remnants are located in the central region of the cluster and form a dark core. We expect that at a few kpc from the Galactic centre, the efficient evaporation of the lower-mass stars caused by the strong tidal force exposes the dark core, because the dynamical properties of the DSC are dominated by the remnants. Due to the invisibility of the remnants, finding a DSC by observation is challenging. In this project, we use $N$-body simulations to obtain models of DSCs and try to discern observables that signify a DSC. We consider four observables: the mass spectrum, the observational mass density profile, the observational velocity dispersion profile and the mass segregation. The models show that a DSC typically exhibits two distinct characteristics: for a given mass in stars and a given half-light radius the expected velocity dispersion is underestimated when only visible stars are considered, and there is a lack of measurable mass segregation among the stars. These properties can be helpful for finding DSCs in observational data, such as the Gaia catalogue.

astro-ph.GA↗

arXiv2024

The spin, expansion and contraction of open star clusters

Empirical constraints on the internal dynamics of open clusters are important for understanding their evolution and evaporation. High precision astrometry from Gaia DR3 are thus useful to observe aspects of the cluster dynamics. This work aims to identify dynamically peculiar clusters such as spinning and expanding clusters. We also quantify the spin frequency and expansion rate and compare them with $N$-body models to identify the origins of the peculiarities. We used the latest Gaia DR3 and archival spectroscopic surveys to analyse the radial velocities and proper motions of the cluster members in 1379 open clusters. A systematic analysis of synthetic clusters is performed to demonstrate the observability of the cluster spin along with effects of observational uncertainties. $N$-body simulations were used to understand the evolution of cluster spin and expansion for initially non-rotating clusters. We identified spin signatures in 10 clusters (and 16 candidates). Additionally, we detected expansion in 18 clusters and contraction in 3 clusters. The expansion rate is compatible with previous theoretical estimates based on expulsion of residual gas. The orientation of the spin axis is independent of the orbital angular momentum. The spin frequencies are much larger than what is expected from simulated initially non-rotating clusters. This indicates that >1% of the clusters are born rotating and/or they have undergone strong interactions. Higher precision observations are required to increase the sample of such dynamically peculiar clusters and to characterise them.

astro-ph.GA↗

arXiv2023

A Novel Estimation Method for Temperature of Magnetic Nanoparticles Dominated by Brownian Relaxation Based on Magnetic Particle Spectroscopy

This paper presents a novel method for estimating the temperature of magnetic nanoparticles (MNPs) based on AC magnetization harmonics of MNPs dominated by Brownian relaxation. The difference in the AC magnetization response and magnetization harmonic between the Fokker-Planck equation and the Langevin function was analyzed, and we studied the relationship between the magnetization harmonic and the key factors, such as Brownian relaxation time, temperature, magnetic field strength, core size and hydrodynamic size of MNPs, excitation frequency, and so on. We proposed a compensation function for AC magnetization harmonic with consideration of the key factors and the difference between the Fokker-Planck equation and the Langevin function. Then a temperature estimation model based on the compensation function and the Langevin function was established. By employing the least squares algorithm, the temperature was successfully calculated. The experimental results show that the temperature error is less than 0.035 K in the temperature range from 310 K to 320 K. The temperature estimation model is expected to improve the performance of the magnetic nanoparticle thermometer and be applied to magnetic nanoparticle-mediated hyperthermia.

physics.med-ph↗

arXiv2023

Machine Learning for the Multi-Dimensional Bin Packing Problem: Literature Review and Empirical Evaluation

The Bin Packing Problem (BPP) is a well-established combinatorial optimization (CO) problem. Since it has many applications in our daily life, e.g. logistics and resource allocation, people are seeking efficient bin packing algorithms. On the other hand, researchers have been making constant advances in machine learning (ML), which is famous for its efficiency. In this article, we first formulate BPP, introducing its variants and practical constraints. Then, a comprehensive survey on ML for multi-dimensional BPP is provided. We further collect some public benchmarks of 3D BPP, and evaluate some online methods on the Cutting Stock Dataset. Finally, we share our perspective on challenges and future directions in BPP. To the best of our knowledge, this is the first systematic review of ML-related methods for BPP.

cs.LG↗

arXiv2022

Measurements of Rayleigh Ratios in Linear Alkylbenzene

In present work an experiment has been designed to measure the Rayleigh ratio directly at $405$ nm and $432$ nm for linear alkylbenzene which is a common solvent used in liquid scintillator detectors of neutrino experiments. The perpendicularly polarized Rayleigh ratio is determined to be $(4.52\pm 0.28)\times 10^{-6}$ m$^{-1}\cdot$ sr$^{-1}$ at $405$ nm and $(3.82\pm 0.24)\times 10^{-6}$ m$^{-1}\cdot$ sr$^{-1}$ at $432$ nm, and the corresponding Rayleigh scattering length is $L_{Ray} = 22.9\pm 0.3(\mathrm{stat.})\pm 1.7(\mathrm{sys.})$ m at $405$ nm and $L_{Ray}= 27.0\pm 0.9(\mathrm{stat.})\pm 1.8(\mathrm{sys.})$ m at $432$ nm.

physics.ins-det↗

arXiv2022

A Monte Carlo Method for Rayleigh Scattering in Liquid Detectors

A new Monte Carlo method has been implemented to describe the angular and polarization distributions of anisotropic liquids, like water and linear alkylbenzene, by considering orientational fluctuations of polarizability tensors. The scattered light of anisotropic liquids is depolarized with an angular distribution of $1+(1-ρ_v)/(1+3ρ_v)\cos^2θ$, which is modified by the depolarization ratio $ρ_v$. A standalone experiment has validated the simulation results of LAB. The new method can provide more accurate knowledge on light propagation in large liquid detectors, which is beneficial to the developments of reconstruction for detectors.

physics.ins-det↗

arXiv2020

Deep-Learning-Based Kinematic Reconstruction for DUNE

In the framework of three-active-neutrino mixing, the charge parity phase, the neutrino mass ordering, and the octant of $θ_{23}$ remain unknown. The Deep Underground Neutrino Experiment (DUNE) is a next-generation long-baseline neutrino oscillation experiment, which aims to address these questions by measuring the oscillation patterns of $ν_μ/ν_e$ and $\barν_μ/\barν_e$ over a range of energies spanning the first and second oscillation maxima. DUNE far detector modules are based on liquid argon TPC (LArTPC) technology. A LArTPC offers excellent spatial resolution, high neutrino detection efficiency, and superb background rejection, while reconstruction in LArTPC is challenging. Deep learning methods, in particular, Convolutional Neural Networks (CNNs), have demonstrated success in classification problems such as particle identification in DUNE and other neutrino experiments. However, reconstruction of neutrino energy and final state particle momenta with deep learning methods is yet to be developed for a full AI-based reconstruction chain. To precisely reconstruct these kinematic characteristics of detected interactions at DUNE, we have developed and will present two CNN-based methods, 2-D and 3-D, for the reconstruction of final state particle direction and energy, as well as neutrino energy. Combining particle masses with the kinetic energy and the direction reconstructed by our work, the four-momentum of final state particles can be obtained. Our models show considerable improvements compared to the traditional methods for both scenarios.

physics.ins-det↗

arXiv2019

Plug-and-Play Priors for Reconstruction-based Placental Image Registration

This paper presents a novel deformable registration framework, leveraging an image prior specified through a denoising function, for severely noise-corrupted placental images. Recent work on plug-and-play (PnP) priors has shown the state-of-the-art performance of reconstruction algorithms under such priors in a range of imaging applications. Integration of powerful image denoisers into advanced registration methods provides our model with a flexibility to accommodate datasets that have low signal-to-noise ratios (SNRs). We demonstrate the performance of our method under a wide variety of denoising models in the context of diffeomorphic image registration. Experimental results show that our model substantially improves the accuracy of spatial alignment in applications of 3D in-utero diffusion-weighted MR images (DW-MRI) that suffer from low SNR and large spatial transformations.

eess.IV↗

arXiv2019

Thermal diffusivity and specific heat capacity of linear alkylbenzene

We report the measurements of the thermal diffusivity and the isobaric specific heat capacity of linear alkylbenzene at about 23$\,^{\circ}\mathrm{C}$ with the standard atmosphere, which are measured for the first time. The conductivity, heat capacity ratio, and speed of sound were derived from our measurements based on thermodynamic relations. The temperature dependence of heat capacity ratio and isobaric specific heat capacity were calculated and relevant results were discussed.

physics.ins-det↗

arXiv2019

A new method of energy reconstruction for large spherical liquid scintillator detectors

Liquid scintillators are commonly used to detect low energy neutrinos from the reactors, sun, and earth. It is a challenge to reconstruct deposited energies for a large liquid scintillator detector. For detectors with multiple optical mediums such as JUNO and SNO+, the prediction of the propagation of detected photons is extremely difficult due to mixed optical processes such as Rayleigh scattering, refraction and total reflection at their boundaries. Calibration based reconstruction methods consume impractical time since a large number of calibration points are required in a giant detector. In this paper, we propose a new model-independent method to reconstruct deposited energies with minimum requirements on the calibration system. This method is validated with JUNO's offline software. Monte Carlo studies show that the energy non-uniformity can be controlled below 1%, which is crucial for JUNO to achieve 3% energy resolution.

physics.ins-det↗

arXiv2016

Occulting Light Concentrators in Liquid Scintillator Neutrino Detectors

The experimental efforts characterizing the era of precision neutrino physics revolve around collecting high-statistics neutrino samples and attaining an excellent energy and position resolution. Next generation liquid-based neutrino detectors, such as JUNO, HyperKamiokande, etc, share the use of a large target mass, and the need of pushing light collection to the edge for maximal calorimetric information. Achieving high light collection implies considerable costs, especially when considering detector masses of several kt. A traditional strategy to maximize the effective photo-coverage with the minimum number of PMTs relies on Light Concentrators (LC), such as Winston Cones. In this paper, the authors introduce a novel concept called Occulting Light Concentrators (OLC), whereby a traditional LC gets tailored to a conventional PMT, by taking into account its single-photoelectron collection efficiency profile and thus occulting the worst performing portion of the photocathode. Thus, the OLC shape optimization takes into account not only the optical interface of the PMT, but also the maximization of the PMT detection performances. The light collection uniformity across the detector is another advantage of the OLC system. By considering the case of JUNO, we will show OLC capabilities in terms of light collection and energy resolution.

physics.ins-det↗

arXiv2016

Test of non-Newtonian gravitational force at micrometer range

We report an experimental test of non-Newtonian gravitational forces at mi- crometer range. To experimentally subtract off the Casimir force and the electrostatic force background, differential force measurements were performed by sensing the lateral force between a gold sphere and a density modulated source mass using a soft cantilever. The current sensitivity is limited by the patch electrostatic force, which is further improved by two dimensional (2D) force mapping. The preliminary result sets a model independent constraint on the Yukawa type force at this range.

gr-qc↗

arXiv2015

Rayleigh scattering and depolarization ratio in linear alkylbenzene

Linear alkylbenzene (LAB) is adopted to be the organic solvent for the Jiangmen Underground Neutrino Observatory (JUNO) liquid scintillator detectors due to the ultra-transparency. However the current Rayleigh scattering length calculation disagrees with the measurement. The present paper for the first time reports the Rayleigh scattering of LAB being anisotropic and the depolarization ratio being 0.31+-0.01(stat.)+-0.01(sys.). We proposed an indirectly method for Rayleigh scattering measurement with Einstein-Smoluchowski-Cabannes formula, and the Rayleigh scattering length of LAB is determined to be 28.2+-1.0 m at 430 nm.

physics.ins-det↗